Fine-Tune GPT-3 to Write an Entire Coherent Novel - Part 1

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!

Grab it

Explore the process of fine-tuning GPT-3 to write an entire coherent novel in this 41-minute video. Learn about generating fiction from a single story premise, copying writing styles, and creating fan fiction generators. Discover techniques for training AI to write stories, generate outlines, and summarize content in real-time. Delve into the intricacies of working memory, task sets, and preprocessing text data for GPT-3. Gain insights on writing novels paragraph by paragraph, managing increasing story lengths, and building narratives chunk by chunk. This comprehensive tutorial covers various aspects of AI-powered novel writing, providing valuable knowledge for those interested in leveraging machine learning for creative writing.

Syllabus

- Fine tuning GPT3 for fiction
- Generating fiction from a single story premise
- GPT3 copying a style verbatim
- Writing a novel: process, what works, and what doesn't
- Creating a fan fiction generator
- Generating fan fiction with a machine
- Training a machine to write a story
- Frankenstein and Alice in Wonderland Outlines
- Generating a story outline with Auto Muse
- The difference between working memory and recall
- The need for a task set in GPT3
- Generating a live summary of a story
- Preprocessing text data for GPT3
- Generating a novel with Auto Muse
- Writing a novel one paragraph at a time
- The increasing length of the story
- The process of creating a machine that can summarize a novel
- Building a story one chunk at a time
- The length of each summary chunk